Load Data and Codes

Read dataset with id, site, period, species, abund

data <- read.csv("data.csv", h = T)
head(data)

##   ID Site Period                   species Nr
## 1  1   42      1            Noctua pronuba  8
## 2  2   42      1      Atlantarctia tigrina  1
## 3  3   42      1           Anarta myrtilli  1
## 4  4   42      1      Lycophotia molothina  7
## 5  5   42      1        Harpyia milhauseri  1
## 6  6   42      1 Pseudoterpna coronillaria  1

library(xtable)
library(boot)
source('chao.R')
source('distanc.R')

Distances

The distances between pairs were calculated using QGIS for IPCC and IGeoE:

Hayford Gauss IPCC:

dist <- read.csv("dist_mat_ipcc.csv", h = T)
dist <- dist[-1]
DIST1 <- distanc(dist)
head(DIST1)

##      [,1] [,2]   [,3]
## [1,]    1    2  60.00
## [2,]    1    3  60.00
## [3,]    1    4  84.85
## [4,]    1    5 247.39
## [5,]    1    6 247.40
## [6,]    1    7 339.41

Hayford Gauss IGeoE

dist <- read.csv("dist_mat_igeoe.csv", h = T)
dist <- dist[-1]
DIST2 <- distanc(dist)
head(DIST2)

##      [,1] [,2]   [,3]
## [1,]    1    2  60.00
## [2,]    1    3  60.00
## [3,]    1    4  84.85
## [4,]    1    5 247.39
## [5,]    1    6 247.40
## [6,]    1    7 339.41

par(mfrow=c(2, 1))
hist(DIST1[, 3], main = "IPCC", xlab = "Geographic distance (m)", ylab = "Frequency")
hist(DIST2[, 3], main = "IGEOE", xlab = "Geographic distance (m)", ylab = "Frequency")

plot of chunk histdistanc

The histograms have the same behavior, then I chose the first vector of distances (IPCC) for analysing the data.

Data

Sorensen and Jaccard Estimators were estimated using unseen species (Chao et al 2005). The data-set was subseted by sample. Variables created from the chao-function have the following columns:

sample1 <- chao(1, 1, "sample1")
sample2 <- chao(2, 2, "sample2")
sample3 <- chao(3, 3, "sample3")
sample4 <- chao(4, 4, "sample4")
sample5 <- chao(5, 5, "sample5")
sample6 <- chao(6, 6, "sample6")

p: sampling period.
s1: the code of sample 1.
s2: the code of sample 2.
n: the number of observed species in the sample 1.
nsp1: the number of species in the sample 1.
m: the number of observed species in the sample 2.
nsp2: the number of species in the sample 2.
nsh: the number of observed shared species in the two samples.
fm1: the observed number of shared species that occur once in sample 2.
fm2: the observed number of shared species that occur twice in sample 2.
f1m: the observed number of shared species that occur once in sample 1.
f2m: the observed number of shared species that occur twice in sample 1.
u_pt1: abundance sum of shared species divided by species number for the sample 1.
v_pt1: abundance sum of shared species divided by species number for the sample 2.
Junad: Jaccard unadjusted estimator.
Lunad: Sorensen unadjusted estimator.
U: u_pt1 plus unseen species for the sample 2.
V: v_pt1 plus unseen species for the sample 1.
Jabd: Jaccard adjusted estimator.
Labd: Sorensen adjusted estimator.

head(sample1)

##      p s1 s2  n nsp1   m nsp2 nsh fm1 fm2 f1m f2m   u_pt1  v_pt1  Junad
## [1,] 1  1  2 44   17  72   26  13   4   2   5   4 0.88636 0.6667 0.6142
## [2,] 1  1  3 44   17  73   33  13   5   1   5   5 0.86364 0.6301 0.5731
## [3,] 1  1  4 44   17   4    4   1   1   1   0   1 0.04545 0.2500 0.0400
## [4,] 1  1  5 44   17 114   33  12   2   1   4   4 0.86364 0.6754 0.6103
## [5,] 1  1  6 44   17  67   25  10   4   2   4   2 0.77273 0.4328 0.3840
## [6,] 1  1  7 44   17  97   29  12   3   1   5   4 0.81818 0.7216 0.6219
##        Lunad      U      V    Jabd   Labd
## [1,] 0.76098 1.0000 0.7854 0.78543 0.8798
## [2,] 0.72864 1.0000 0.7172 0.71715 0.8353
## [3,] 0.07692 0.0625 0.2500 0.05263 0.1000
## [4,] 0.75803 0.9312 0.7440 0.70526 0.8272
## [5,] 0.55487 1.0000 0.5495 0.54953 0.7093
## [6,] 0.76689 1.0000 0.9231 0.92315 0.9600

Exploratory Analysis

These functions were used to synthesize the samples for each year. Moth samplings were conducted for two years (2011 and 2012), each year had three sampling period.

trisamples <- function(base1, base2, base3, vr){
  base <- matrix(0, nrow = dim(base1)[1])
  for(i in 1:dim(base1)[1]){
    base[i] <- (base1[i, vr] + base2[i, vr] + base3[i, vr])/3
  }
  base
}

bisamples <- function(base1, base2){
  base <- matrix(0, nrow = length(base1))
  for(i in 1:length(base1)){
    base[i] <- (base1[i] + base2[i])/2
  }
  base
}

Calculated the dissimilarity average per site for first year

Jperiod1 <- trisamples(sample1, sample2, sample3, 19)
Lperiod1 <- trisamples(sample1, sample2, sample3, 20)

par(mfrow=c(1, 2))
plot(DIST1[, 3], (1 - Jperiod1), ylim = c(0, 1), xlab = "Geographic Distance (m)", ylab = 
       expression(~beta*"-diversity"), main = "Jaccard")
plot(DIST1[, 3], (1 - Lperiod1), ylim = c(0, 1), xlab = "Geographic Distance (m)", ylab = 
       expression(~beta*"-diversity"), main = "Sorensen")

plot of chunk firstyear

Calculated the dissimilarity average per site for second year

Jperiod2 <- trisamples(sample4, sample5, sample6, 19)
Lperiod2 <- trisamples(sample4, sample5, sample6, 20)

par(mfrow=c(1, 2))
plot(DIST1[, 3], (1 - Jperiod2), ylim = c(0, 1), xlab = "Geographic Distance (m)", ylab = 
       expression(~beta*"-diversity"), main = "Jaccard")
plot(DIST1[, 3], (1 - Lperiod2), ylim = c(0, 1), xlab = "Geographic Distance (m)", ylab = 
       expression(~beta*"-diversity"), main = "Sorensen")

plot of chunk secondyear

Calculated the dissimilarity average per site for both years

Jperiod <- bisamples(Jperiod1, Jperiod2)
Lperiod <- bisamples(Lperiod1, Lperiod2)

par(mfrow=c(1, 2))
plot(DIST1[, 3], (1 - Jperiod), ylim = c(0, 1), xlab = "Geographic Distance (m)", ylab = 
       expression(~beta*"-diversity"), main = "Jaccard")
plot(DIST1[, 3], (1 - Lperiod), ylim = c(0, 1), xlab = "Geographic Distance (m)", ylab = 
       expression(~beta*"-diversity"), main = "Sorensen")

plot of chunk allyears

DATASET: I created a new database with six columns:

p: no functionality, I created just to standardize with other functions previously created.
s1: site 1
s2: site 2
Jabd: Jaccard adjusted estimator
Labd: Sorensen adjusted estimator
ipcc: distance between site 1 and site 2 (Hayford Gauss IPCC)

b.div.mean <- cbind(sample1[, 1], sample1[, 2], sample1[, 3], Jperiod, Lperiod, DIST1[, 3])
colnames(b.div.mean) <- c("p", "s1", "s2", "Jabd", "Labd", "ipcc")

head(b.div.mean)

##      p s1 s2   Jabd   Labd   ipcc
## [1,] 1  1  2 0.6520 0.7724  60.00
## [2,] 1  1  3 0.7169 0.8153  60.00
## [3,] 1  1  4 0.2347 0.3054  84.85
## [4,] 1  1  5 0.5494 0.6839 247.39
## [5,] 1  1  6 0.6330 0.7575 247.40
## [6,] 1  1  7 0.6073 0.7244 339.41

Samples were taken in three landscapes that represent the farmland abandonment gradient: meadow-dominated, shrub-dominated and forest-dominated. In each landscape had 28 fixed sampling sites divided into four biotopes (meadow, short shrub, tall shrub and woodland), totaling 84 sampling sites.

forest-dominated:

mf <- c(56) #m - meadow
sf <- c(36, 39, 42, 45, 46) #s - short shrub
tf <- c(34, 49) #t - tall shrub
wf <- c(29, 30, 31, 32, 33, 35, 37, 38, 40, 41, 43, 44, 47, 48, 50, 51, 52, 53, 54, 55) 
#w - woodland

shrub-dominated (m - mixed)

mm <- c(68, 75, 77, 84)
sm <- c(60, 64, 65, 66, 67, 70, 78, 79, 82)
tm <- c(57, 58, 59, 61, 62, 63, 69, 71, 72, 74, 76)
wm <- c(73, 80, 81, 83)

meadow-dominated (a - agriculture)

ma <- c(1, 9, 14, 15, 16, 17, 18, 19, 21, 23, 24, 25, 27)
sa <- c(2, 8, 10, 11, 12, 13, 20)
ta <- c(3, 4, 5, 6, 7, 28)
wa <- c(22, 26)

FIRST APROARCH

Within landscape

landsc_within <- function(base, s1, s2){
  result <- NULL
  for(i in s1){
    for(j in s2){
      if(i != j){
        result <- rbind(result, base[base[, 2] == i & base[, 3] == j, ])
      }
    }
  }
  result
}

idmea <- c(ma, sa, ta, wa)
idmix <- c(mm, sm, tm, wm)
idfor <- c(mf, sf, tf, wf)

m.mea <- landsc_within(b.div.mean, idmea, idmea)
m.mix <- landsc_within(b.div.mean, idmix, idmix)
m.for <- landsc_within(b.div.mean, idfor, idfor)

This function creates a database for each landscape, but with all pairs within landscape.

Regression models

Jaccard Estimator

jm.modelfor <- lm((1 - m.for[, 4]) ~ m.for[, 6])
jm.modelmix <- lm((1 - m.mix[, 4]) ~ m.mix[, 6])
jm.modelmea <- lm((1 - m.mea[, 4]) ~ m.mea[, 6])

Sorensen Estimator

sm.modelfor <- lm((1 - m.for[, 5]) ~ m.fordata[, 6])
sm.modelmix <- lm((1 - m.mix[, 5]) ~ m.mixdata[, 6])
sm.modelmea <- lm((1 - m.mea[, 5]) ~ m.meadata[, 6])

Forest

plot of chunk forestlandscape1

Shrub

plot of chunk shrublandscape1

Meadow

plot of chunk meadowlandscape1

Mean Dissimilarity for each scale

plot.scale2 <- function(base, estimator = 0){
  #if estimator == 0: Jaccard, otherwise: Sorensen (default is Jaccard)
  area <- base[, 6]
  if(estimator) div <- base[, 5]
  else div <- base[, 4]
  plot(area, 1 - div, ylim = c(0, 1), xlim = c(0, 2000), xlab = "Geographic Distance (m)", 
       ylab = expression(~beta*"-diversity"))
  abline(v = c(115, 455), col = "gray", lwd = 1.5, lty = 2)
  lines(1:115, rep(1 - mean(div[area <= 115]), 115), col = 2, lwd = 4.5)
  lines(116:445, rep(1 - mean(div[(area <= 445) & (area > 115)]), 330), col = 2, lwd = 4.5)
  lines(446:1810, rep(1 - mean(div[area > 445]), 1365), col = 2, lwd = 4.5)
}

Forest

par(mfrow = c(1, 2))
plot.scale2(m.for)
plot.scale2(m.for, 1)

plot of chunk forestscale

Shrub

plot of chunk shrubscale

Meadow

plot of chunk meadowscale

Bootstrapping

# function to obtain regression weights 
bs.jac <- function(formula, data, ind){
  d <- data[ind,] # allows boot to select sample
  fit <- lm(formula, data = d)
  return(c(coef(fit), summary(fit)$r.square, mean(d[, 4]))) 
} 

bs.sor <- function(formula, data, ind){
  d <- data[ind,] # allows boot to select sample
  fit <- lm(formula, data = d)
  return(c(coef(fit), summary(fit)$r.square, mean(d[, 5]))) 
}

Within landscape

# bootstrapping with 2000 replications
jforres <- boot(data = as.data.frame(m.fordata), statistic = bs.jac, R = 2000, 
                formula = (1 - Jabd) ~ ipcc)
sforres <- boot(data = as.data.frame(m.fordata), statistic = bs.sor, R = 2000, 
                formula = (1 - Labd) ~ ipcc)
jmixres <- boot(data = as.data.frame(m.mixdata), statistic = bs.jac, R = 2000, 
                formula = (1 - Jabd) ~ ipcc)
smixres <- boot(data = as.data.frame(m.mixdata), statistic = bs.sor, R = 2000, 
                formula = (1 - Labd) ~ ipcc)

jmeares <- boot(data = as.data.frame(m.meadata), statistic = bs.jac, R = 2000, 
                formula = (1 - Jabd) ~ ipcc)
smeares <- boot(data = as.data.frame(m.meadata), statistic = bs.sor, R = 2000, 
                formula = (1 - Labd) ~ ipcc)

Confidence Interval

Bootstrap 95% CI for regression coefficients

Jaccard

jforres.ci <- ci.boot(jforres)
jmixres.ci <- ci.boot(jmixres)
jmeares.ci <- ci.boot(jmeares)

Sorensen

jforres.ci <- ci.boot(jforres)
jmixres.ci <- ci.boot(jmixres)
jmeares.ci <- ci.boot(jmeares)

Results

Intercept

par(mfrow = c(1, 2))
interc.plot3(jforres.ci, jmixres.ci, jmeares.ci)
title("Jaccard", cex = 1.5)
interc.plot3(sforres.ci, smixres.ci, smeares.ci)
title("Sorensen", cex = 1.5)

plot of chunk intercept_meth1

Dissimilarity

par(mfrow = c(1, 2))
dissim.plot3(jforres.ci, jmixres.ci, jmeares.ci)
title("Jaccard", cex = 1.5)
dissim.plot3(sforres.ci, smixres.ci, smeares.ci)
title("Sorensen", cex = 1.5)

plot of chunk dissim_meth1

Slope

par(mfrow = c(1, 2))
slope.plot3(jforres.ci, jmixres.ci, jmeares.ci)
title("Jaccard", cex = 1.5)
slope.plot3(sforres.ci, smixres.ci, smeares.ci)
title("Sorensen", cex = 1.5)

plot of chunk slope_meth1

R-Square

par(mfrow = c(1, 2))
rsquar.plot3(jforres.ci, jmixres.ci, jmeares.ci)
title("Jaccard", cex = 1.5)
rsquar.plot3(sforres.ci, smixres.ci, smeares.ci)
title("Sorensen", cex = 1.5)

plot of chunk r-squar_meth1

Tables

Jaccard

print(xtable(boot.table2(jforres.ci, jmixres.ci, jmeares.ci), 
             digits = 4, caption = "Intercept"), type = "latex", comment = FALSE)

Sorensen

print(xtable(boot.table2(sforres.ci, smixres.ci, smeares.ci), 
             digits = 4, caption = "Intercept"), type = "latex", comment = FALSE)

SECOND APPROACH

within landscapes

agr.mea <- landsc_within(b.div.mean, ma, ma)
shr.shr <- landsc_within(b.div.mean, sm, sm)
tal.shr <- landsc_within(b.div.mean, tm, tm)
for.woo <- landsc_within(b.div.mean, wf, wf)
mix.shr <- landsc_within(b.div.mean, c(sm, tm), c(sm, tm))

This function creates a database for each landscape. Also, I analyzed separately short and tall shrub within the shrub-dominated landscape.

In this analysis only the biotopes that correspond the type of the landscape were considered.

Regression models

Jaccard Estimator

jmodelfor.woo <- lm((1 - for.woo[, 4]) ~ for.woo[, 6])
jmodelmix.shr <- lm((1 - mix.shr[, 4]) ~ mix.shr[, 6])
jmodelagr.mea <- lm((1 - agr.mea[, 4]) ~ agr.mea[, 6])
jmodelshr.shr <- lm((1 - shr.shr[, 4]) ~ shr.shr[, 6])
jmodeltal.shr <- lm((1 - tal.shr[, 4]) ~ tal.shr[, 6])

Sorensen Estimator

smodelfor.woo <- lm((1 - for.woo[, 5]) ~ for.woo[, 6])
smodelmix.shr <- lm((1 - mix.shr[, 5]) ~ mix.shr[, 6])
smodelagr.mea <- lm((1 - agr.mea[, 5]) ~ agr.mea[, 6])
smodelshr.shr <- lm((1 - shr.shr[, 5]) ~ shr.shr[, 6])
smodeltal.shr <- lm((1 - tal.shr[, 5]) ~ tal.shr[, 6])

Short and tall shrubs are together in Shrub (right side: Jaccard and left side: Sorensen)

Forest

plot of chunk forest_within

Shrub

plot of chunk shrub_within

Meadow

plot of chunk meadow_within

Short and tall shrubs are separated (right side: Jaccard and left side: Sorensen)

Tall Shrub

plot of chunk tall_within

Short Shrub

plot of chunk short_within

Mean Dissimilaritiy for each landscape (with standard deviations)

par(mfrow=c(1, 2))
m.data <- c(mean(for.woo[, 4]), mean(mix.shr[, 4]), mean(agr.mea[, 4]))
m.li <- c(mean(for.woo[, 4]) - sd(for.woo[, 4]), mean(mix.shr[, 4]) - sd(mix.shr[, 4]), 
          mean(agr.mea[, 4]) - sd(agr.mea[, 4]))
m.lu <- c(mean(for.woo[, 4]) + sd(for.woo[, 4]), mean(mix.shr[, 4]) + sd(mix.shr[, 4]), 
          mean(agr.mea[, 4]) + sd(agr.mea[, 4]))
plot(1:3, 1 - m.data, xaxt = "n", ylim=c(0, 1), xlab = "", 
     ylab = expression(~beta*"-diversity"), main = "Jaccard adjusted")
axis(1, at = 1:3, labels = c("Forest", "Shrub", "Meadow"))
segments(x0 = 1:3, y0 = 1 - m.li, y1 = 1 - m.lu, col = 1)

m.data <- c( mean(for.woo[, 5]), mean(mix.shr[, 5]), mean(agr.mea[, 5]))
m.li <- c(mean(for.woo[, 5]) - sd(for.woo[, 5]), mean(mix.shr[, 5]) - sd(mix.shr[, 5]), 
          mean(agr.mea[, 5]) - sd(agr.mea[, 5]))
m.lu <- c(mean(for.woo[, 5]) + sd(for.woo[, 5]), mean(mix.shr[, 5]) + sd(mix.shr[, 5]), 
          mean(agr.mea[, 5]) + sd(agr.mea[, 5]))
plot(1:3, 1 - m.data, xaxt = "n", ylim=c(0, 1), xlab = "", 
     ylab = expression(~beta*"-diversity"), main = "Sorensen adjusted")
axis(1, at = 1:3, labels = c("Forest", "Shrub", "Meadow"))
segments(x0 = 1:3, y0 = 1 - m.li, y1 = 1 - m.lu, col = 1)

plot of chunk dissim3_within

Mean Dissimilarity for each landscape (with tall and short shrub separately)

par(mfrow=c(1, 2))
m.data <- c(mean(for.woo[, 4]), mean(tal.shr[, 4]), mean(shr.shr[, 4]), mean(agr.mea[, 4]))
m.li <- c(mean(for.woo[, 4]) - sd(for.woo[, 4]), mean(tal.shr[, 4]) - sd(tal.shr[, 4]), 
          mean(shr.shr[, 4]) - sd(shr.shr[, 4]), mean(agr.mea[, 4]) - sd(agr.mea[, 4]))
m.lu <- c(mean(for.woo[, 4]) + sd(for.woo[, 4]), mean(tal.shr[, 4]) + sd(tal.shr[, 4]), 
          mean(shr.shr[, 4]) + sd(shr.shr[, 4]), mean(agr.mea[, 4]) + sd(agr.mea[, 4]))
plot(1:4, 1 - m.data, xaxt = "n", ylim=c(0, 1), xlab = "", ylab = 
       expression(~beta*"-diversity"), main = "Jaccard adjusted")
axis(1, at = 1:4, labels = c("Forest", "Tall", "Short", "Meadow"))
segments(x0 = 1:4, y0 = 1 - m.li, y1 = 1 - m.lu, col = 1)

m.data <- c(mean(for.woo[, 5]), mean(tal.shr[, 5]), mean(shr.shr[, 5]), mean(agr.mea[, 5]))
m.li <- c(mean(for.woo[, 5]) - sd(for.woo[, 5]), mean(tal.shr[, 5]) - sd(tal.shr[, 5]), 
          mean(shr.shr[, 5]) - sd(shr.shr[, 5]), mean(agr.mea[, 5]) - sd(agr.mea[, 5]))
m.lu <- c(mean(for.woo[, 5]) + sd(for.woo[, 5]), mean(tal.shr[, 5]) + sd(tal.shr[, 5]), 
          mean(shr.shr[, 5]) + sd(shr.shr[, 5]), mean(agr.mea[, 5]) + sd(agr.mea[, 5]))
plot(1:4, 1 - m.data, xaxt = "n", ylim=c(0, 1), xlab = "", 
     ylab = expression(~beta*"-diversity"), main = "Sorensen adjusted")
axis(1, at = 1:4, labels = c("Forest", "Tall", "Short", "Meadow"))
segments(x0 = 1:4, y0 = 1 - m.li, y1 = 1 - m.lu, col = 1)

plot of chunk dissim4_within

between landscapes

landsc_between <- function(base, s1, s2){
  result <- NULL
  for(i in s1){
    for(j in s2){
      result <- rbind(result, base[(base[, 2] == i & base[, 3] == j) | 
                                     (base[, 2] == j & base[, 3] == i), ])
    }
  }
  result
}

This function creates a database of pairs of biotypes of the same type, but are in different landscapes. Also, I analyzed separately short and tall shrub among landscapes.

mea <- landsc_between(b.div.mean, ma, c(mm, mf))
shr <- landsc_between(b.div.mean, sm, c(sa, sf))
tal <- landsc_between(b.div.mean, tm, c(ta, tf))
woo <- landsc_between(b.div.mean, wf, c(wa, wm))
mix <- rbind(shr, tal)

Regreesion models

Jaccard Estimator

jmodelwoo <- lm((1 - woo[, 4]) ~ woo[, 6])
jmodelmix <- lm((1 - mix[, 4]) ~ mix[, 6])
jmodelmea <- lm((1 - mea[, 4]) ~ mea[, 6])
jmodelshr <- lm((1 - shr[, 4]) ~ shr[, 6])
jmodeltal <- lm((1 - tal[, 4]) ~ tal[, 6])

Sorensen Estimator

smodelwoo <- lm((1 - woo[, 5]) ~ woo[, 6])
smodelmix <- lm((1 - mix[, 5]) ~ mix[, 6])
smodelmea <- lm((1 - mea[, 5]) ~ mea[, 6])
smodelshr <- lm((1 - shr[, 5]) ~ shr[, 6])
smodeltal <- lm((1 - tal[, 5]) ~ tal[, 6])

Short and tall shrubs are together in Shrub (right side: Jaccard and left side: Sorensen)

Forest

plot of chunk forest_between

Shrub

plot of chunk shrub_between

Meadow

plot of chunk meadow_between

Short and tall shrubs are separated (right side: Jaccard and left side: Sorensen)

Tall Shrub

plot of chunk tall_between

Short Shrub

plot of chunk short_between

Mean Dissimilarity for each landscape (with standard deviations)

par(mfrow=c(1, 2))
m.data <- c(mean(woo[, 4]), mean(mix[, 4]), mean(mea[, 4]) )
m.li <- c(mean(woo[, 4]) - sd(woo[, 4]), mean(mix[, 4]) - sd(mix[, 4]), 
          mean(mea[, 4]) - sd(mea[, 4]))
m.lu <- c(mean(woo[, 4]) + sd(woo[, 4]), mean(mix[, 4]) + sd(mix[, 4]), 
          mean(mea[, 4]) + sd(mea[, 4]))
plot(1:3, 1 - m.data, xaxt = "n", ylim=c(0, 1), xlab = "", 
     ylab = expression(~beta*"-diversity"), main = "Jaccard adjusted")
axis(1, at = 1:3, labels = c("Forest", "Shrub", "Meadow"))
segments(x0 = 1:3, y0 = 1 - m.li, y1 = 1 - m.lu, col = 1)

m.data <- c(mean(woo[, 5]), mean(mix[, 5]), mean(mea[, 5]) )
m.li <- c(mean(woo[, 5]) - sd(woo[, 5]), mean(mix[, 5]) - sd(mix[, 5]), 
          mean(mea[, 5]) - sd(mea[, 5]))
m.lu <- c(mean(woo[, 5]) + sd(woo[, 5]), mean(mix[, 5]) + sd(mix[, 5]), 
          mean(mea[, 5]) + sd(mea[, 5]))
plot(1:3, 1 - m.data, xaxt = "n", ylim=c(0, 1), xlab = "", 
     ylab = expression(~beta*"-diversity"), main = "Sorensen adjusted")
axis(1, at = 1:3, labels = c("Forest", "Shrub", "Meadow"))
segments(x0 = 1:3, y0 = 1 - m.li, y1 = 1 - m.lu, col = 1)

plot of chunk dissim3_between

Mean Dissimilarity for each landscape (with tall and short shrub separately)

par(mfrow=c(1, 2))
m.data <- c(mean(woo[, 4]), mean(tal[, 4]), mean(shr[, 4]), mean(mea[, 4]) )
m.li <- c(mean(woo[, 4]) - sd(woo[, 4]), mean(tal[, 4]) - sd(tal[, 4]), 
          mean(shr[, 4]) - sd(shr[, 4]), mean(mea[, 4]) - sd(mea[, 4]))
m.lu <- c(mean(woo[, 4]) + sd(woo[, 4]), mean(tal[, 4]) + sd(tal[, 4]), 
          mean(shr[, 4]) + sd(shr[, 4]), mean(mea[, 4]) + sd(mea[, 4]))
plot(1:4, 1 - m.data, xaxt = "n", ylim=c(0, 1), xlab = "", 
     ylab = expression(~beta*"-diversity"), main = "Jaccard adjusted")
axis(1, at = 1:4, labels = c("Forest", "Tall", "Short", "Meadow"))
segments(x0 = 1:4, y0 = 1 - m.li, y1 = 1 - m.lu, col = 1)

m.data <- c(mean(woo[, 5]), mean(tal[, 5]), mean(shr[, 5]), mean(mea[, 5]) )
m.li <- c(mean(woo[, 5]) - sd(woo[, 5]), mean(tal[, 5]) - sd(tal[, 5]), 
          mean(shr[, 5]) - sd(shr[, 5]), mean(mea[, 5]) - sd(mea[, 5]))
m.lu <- c(mean(woo[, 5]) + sd(woo[, 5]), mean(tal[, 5]) + sd(tal[, 5]), 
          mean(shr[, 5]) + sd(shr[, 5]), mean(mea[, 5]) + sd(mea[, 5]))
plot(1:4, 1 - m.data, xaxt = "n", ylim=c(0, 1), xlab = "", 
     ylab = expression(~beta*"-diversity"), main = "Sorensen adjusted")
axis(1, at = 1:4, labels = c("Forest", "Tall", "Short", "Meadow"))
segments(x0 = 1:4, y0 = 1 - m.li, y1 = 1 - m.lu, col = 1)

plot of chunk dissim4_between

Biotopes

mfma <- c(mf, mm, ma)
sfma <- c(sf, sm, sa)
tfma <- c(tf, tm, ta)
wfma <- c(wf, wm, wa)
mixf <- c(sf, sm, sa, tf, tm, ta)

This function creates databse for each biotope independent landscape.

mea.bio <- landsc_within(b.div.mean, mfma, mfma)
sho.bio <- landsc_within(b.div.mean, sfma, sfma)
tal.bio <- landsc_within(b.div.mean, tfma, tfma)
woo.bio <- landsc_within(b.div.mean, wfma, wfma)
mix.bio <- landsc_within(b.div.mean, mixf, mixf)

Regression models

Jaccard Estimator

jmodelwoo.bio <- lm((1 - woo.bio[, 4]) ~ woo.bio[, 6])
jmodelmix.bio <- lm((1 - mix.bio[, 4]) ~ mix.bio[, 6])
jmodelmea.bio <- lm((1 - mea.bio[, 4]) ~ mea.bio[, 6])
jmodelsho.bio <- lm((1 - sho.bio[, 4]) ~ sho.bio[, 6])
jmodeltal.bio <- lm((1 - tal.bio[, 4]) ~ tal.bio[, 6])

Sorensen Estimator

smodelwoo.bio <- lm((1 - woo.bio[, 5]) ~ woo.bio[, 6])
smodelmix.bio <- lm((1 - mix.bio[, 5]) ~ mix.bio[, 6])
smodelmea.bio <- lm((1 - mea.bio[, 5]) ~ mea.bio[, 6])
smodelsho.bio <- lm((1 - sho.bio[, 5]) ~ sho.bio[, 6])
smodeltal.bio <- lm((1 - tal.bio[, 5]) ~ tal.bio[, 6])

Short and tall shrubs are together (right side: Jaccard and left side: Sorensen)

Forest

plot of chunk forest_all

Shrub

plot of chunk shrub_all

Meadow

plot of chunk meadow_all

Short and tall shrubs are separated (right side: Jaccard and left side: Sorensen)

Tall Shrub

plot of chunk tall_all

Short Shrub

plot of chunk short_all

Mean Dissimilarity for each landscape (with standard deviations)

par(mfrow=c(1, 2))
m.data <- c(mean(woo.bio[, 4]), mean(mix.bio[, 4]), mean(mea.bio[, 4]))
m.li <- c(mean(woo.bio[, 4]) - sd(woo.bio[, 4]), mean(mix.bio[, 4]) - sd(mix.bio[, 4]), 
          mean(mea.bio[, 4]) - sd(mea.bio[, 4]))
m.lu <- c(mean(woo.bio[, 4]) + sd(woo.bio[, 4]), mean(mix.bio[, 4]) + sd(mix.bio[, 4]), 
          mean(mea.bio[, 4]) + sd(mea.bio[, 4]))
plot(1:3, 1 - m.data, xaxt = "n", ylim=c(0, 1), xlab = "", 
     ylab = expression(~beta*"-diversity"), main = "Jaccard adjusted")
axis(1, at = 1:3, labels = c("Forest", "Shrub", "Meadow"))
segments(x0 = 1:3, y0 = 1 - m.li, y1 = 1 - m.lu, col = 1)

m.data <- c( mean(woo.bio[, 5]), mean(mix.bio[, 5]), mean(mea.bio[, 5]))
m.li <- c(mean(woo.bio[, 5]) - sd(woo.bio[, 5]), mean(mix.bio[, 5]) - sd(mix.bio[, 5]), 
          mean(mea.bio[, 5]) - sd(mea.bio[, 5]))
m.lu <- c(mean(woo.bio[, 5]) + sd(woo.bio[, 5]), mean(mix.bio[, 5]) + sd(mix.bio[, 5]), 
          mean(mea.bio[, 5]) + sd(mea.bio[, 5]))
plot(1:3, 1 - m.data, xaxt = "n", ylim=c(0, 1), xlab = "", 
     ylab = expression(~beta*"-diversity"), main = "Sorensen adjusted")
axis(1, at = 1:3, labels = c("Forest", "Shrub", "Meadow"))
segments(x0 = 1:3, y0 = 1 - m.li, y1 = 1 - m.lu, col = 1)

plot of chunk dissim3_all

Mean Dissimilarity for each landscape (with tall and short shrubs separately)

par(mfrow=c(1, 2))
m.data <- c(mean(woo.bio[, 4]), mean(tal.bio[, 4]), mean(sho.bio[, 4]), mean(mea.bio[, 4]))
m.li <- c(mean(woo.bio[, 4]) - sd(woo.bio[, 4]), mean(tal.bio[, 4]) - sd(tal.bio[, 4]), 
          mean(sho.bio[, 4]) - sd(sho.bio[, 4]), mean(mea.bio[, 4]) - sd(mea.bio[, 4]))
m.lu <- c(mean(woo.bio[, 4]) + sd(woo.bio[, 4]), mean(tal.bio[, 4]) + sd(tal.bio[, 4]), 
          mean(sho.bio[, 4]) + sd(sho.bio[, 4]), mean(mea.bio[, 4]) + sd(mea.bio[, 4]))
plot(1:4, 1 - m.data, xaxt = "n", ylim=c(0, 1), xlab = "", 
     ylab = expression(~beta*"-diversity"), main = "Jaccard adjusted")
axis(1, at = 1:4, labels = c("Forest", "Tall", "Short", "Meadow"))
segments(x0 = 1:4, y0 = 1 - m.li, y1 = 1 - m.lu, col = 1)

m.data <- c(mean(woo.bio[, 5]), mean(tal.bio[, 5]), mean(sho.bio[, 5]), mean(mea.bio[, 5]))
m.li <- c(mean(woo.bio[, 5]) - sd(woo.bio[, 5]), mean(tal.bio[, 5]) - sd(tal.bio[, 5]), 
          mean(sho.bio[, 5]) - sd(sho.bio[, 5]), mean(mea.bio[, 5]) - sd(mea.bio[, 5]))
m.lu <- c(mean(woo.bio[, 5]) + sd(woo.bio[, 5]), mean(tal.bio[, 5]) + sd(tal.bio[, 5]), 
          mean(sho.bio[, 5]) + sd(sho.bio[, 5]), mean(mea.bio[, 5]) + sd(mea.bio[, 5]))
plot(1:4, 1 - m.data, xaxt = "n", ylim=c(0, 1), xlab = "", 
     ylab = expression(~beta*"-diversity"), main = "Sorensen adjusted")
axis(1, at = 1:4, labels = c("Forest", "Tall", "Short", "Meadow"))
segments(x0 = 1:4, y0 = 1 - m.li, y1 = 1 - m.lu, col = 1)

plot of chunk dissim4_all

Mean Dissimilarity for each scale

plot.scale <- function(base, estimator = 0){
  #if estimator == 0: Jaccard, otherwise: Sorensen (default is Jaccard)
  area <- base[, 6]
  if(estimator) div <- base[, 5]
  else div <- base[, 4]
  plot(area, 1 - div, ylim = c(0, 1), xlim = c(0, 7500), xlab = "Geographic Distance (m)", 
       ylab = expression(~beta*"-diversity"))
  abline(v = c(115, 455, 1810), col = "gray", lwd = 1.5, lty = 2)
  lines(1:115, rep(1 - mean(div[area <= 115]), 115), col = 2, lwd = 4.5)
  lines(116:445, rep(1 - mean(div[(area <= 445) & (area > 115)]), 330), col = 2, lwd = 4.5)
  lines(446:1810, rep(1 - mean(div[(area <= 1810) & (area > 445)]), 1365), col = 2, lwd = 4.5)
  lines(1811:7500, rep(1 - mean(div[area > 1811]), 5690), col = 2, lwd = 4.5)
}

Forest

par(mfrow=c(1, 2))
plot.scale(woo.bio)
plot.scale(woo.bio, 1)

plot of chunk forest_scale

Shrub

plot of chunk shrub_scale

Meadow

plot of chunk meadow_scale

Mean Dissimilarity for each scale (with tall and short shrubs separately)

Tall Shrub

plot of chunk tall_scale

Short Shrub

plot of chunk short_scale

Bootstrapping 1

Mean Dissimilarity for each scale

mean.fun <- function(data, ind){
  mean(data[ind], na.rm = TRUE)
}

plot.boot <- function(base, ci.result, estimator = 0){
  #if estimator == 0: Jaccard, otherwise: Sorensen (default is Jaccard)
  area <- base[, 6]
  if(estimator) div <- base[, 5]
  else div <- base[, 4]
  plot(area, 1 - div, ylim = c(0, 1), xlim = c(0, 7500), xlab = "Geographic Distance (m)", 
       ylab = expression(~beta*"-diversity"))
  abline(v = c(115, 455, 1810), col = "gray", lwd = 1.5, lty = 2)
  lines(1:115, rep(ci.result[1, 1], 115), col = 2, lwd = 4.5)
  lines(116:445, rep(ci.result[2, 1], 330), col = 2, lwd = 4.5)
  lines(446:1810, rep(ci.result[3, 1], 1365), col = 2, lwd = 4.5)
  lines(1811:7500, rep(ci.result[4, 1], 5690), col = 2, lwd = 4.5)
  lines(1:115, rep(ci.result[1, 2], 115), col = 3, lwd = 4.5)
  lines(116:445, rep(ci.result[2, 2], 330), col = 3, lwd = 4.5)
  lines(446:1810, rep(ci.result[3, 2], 1365), col = 3, lwd = 4.5)
  lines(1811:7500, rep(ci.result[4, 2], 5690), col = 3, lwd = 4.5)
  lines(1:115, rep(ci.result[1, 3], 115), col = 3, lwd = 4.5)
  lines(116:445, rep(ci.result[2, 3], 330), col = 3, lwd = 4.5)
  lines(446:1810, rep(ci.result[3, 3], 1365), col = 3, lwd = 4.5)
  lines(1811:7500, rep(ci.result[4, 3], 5690), col = 3, lwd = 4.5)
}

Forest

jsf <- boot(data = 1 - woo.bio[woo.bio[, 6] <= 115, 4], statistic = mean.fun, R = 2000)
jmf <- boot(data = 1 - woo.bio[(woo.bio[, 6] <= 445) & (woo.bio[, 6] > 115), 4], 
            statistic = mean.fun, R = 2000)
jlf <- boot(data = 1 - woo.bio[(woo.bio[, 6] <= 1810) & (woo.bio[, 6] > 445), 4], 
            statistic = mean.fun, R = 2000)
jxf <- boot(data = 1 - woo.bio[woo.bio[, 6] > 1811, 4], statistic = mean.fun, R = 2000)

ci.jsf <- boot.ci(jsf, type = "bca")
ci.jmf <- boot.ci(jmf, type = "bca")
ci.jlf <- boot.ci(jlf, type = "bca")
ci.jxf <- boot.ci(jxf, type = "bca")

jci.for <- rbind(cbind(ci.jsf$t0, ci.jsf$bca[4], ci.jsf$bca[5]), cbind(ci.jmf$t0, 
           ci.jmf$bca[4], ci.jmf$bca[5]), cbind(ci.jlf$t0, ci.jlf$bca[4], 
           ci.jlf$bca[5]), cbind(ci.jxf$t0, ci.jxf$bca[4], ci.jxf$bca[5]))
colnames(jci.for) <- c("mean", "lower.ci", "upper.ci")
rownames(jci.for) <- c("small", "medium", "large", "extra-large")

ssf <- boot(data = 1 - woo.bio[woo.bio[, 6] <= 115, 5], statistic = mean.fun, R = 2000)
smf <- boot(data = 1 - woo.bio[(woo.bio[, 6] <= 445) & (woo.bio[, 6] > 115), 5], 
            statistic = mean.fun, R = 2000)
slf <- boot(data = 1 - woo.bio[(woo.bio[, 6] <= 1810) & (woo.bio[, 6] > 445), 5],
            statistic = mean.fun, R = 2000)
sxf <- boot(data = 1 - woo.bio[woo.bio[, 6] > 1811, 5], statistic = mean.fun, R = 2000)

ci.ssf <- boot.ci(ssf, type = "bca")
ci.smf <- boot.ci(smf, type = "bca")
ci.slf <- boot.ci(slf, type = "bca")
ci.sxf <- boot.ci(sxf, type = "bca")

sci.for <- rbind(cbind(ci.ssf$t0, ci.ssf$bca[4], ci.ssf$bca[5]), cbind(ci.smf$t0, 
           ci.smf$bca[4], ci.smf$bca[5]), cbind(ci.slf$t0, ci.slf$bca[4], 
          ci.slf$bca[5]), cbind(ci.sxf$t0, ci.sxf$bca[4], ci.sxf$bca[5]))
colnames(sci.for) <- c("mean", "lower.ci", "upper.ci")
rownames(sci.for) <- c("small", "medium", "large", "extra-large")

par(mfrow=c(1, 2))
plot.boot(woo.bio, jci.for)
plot.boot(woo.bio, sci.for, 1)

plot of chunk forestscale_boot

Shrub

plot of chunk shrubscale_boot

Meadow

plot of chunk meadowscale_boot

Mean Dissimilarity for each scale (with tall and short shrubs separately)

Tall Shrub

plot of chunk tallscale_boot

Short Shrub

plot of chunk shortscale_boot

Bootstrapping 2

Within Landscape

# bootstrapping with 2000 replications
wl.jforres <- boot(data = as.data.frame(for.woo), statistic = bs.jac, R = 2000, 
                   formula = (1 - Jabd) ~ ipcc)
wl.sforres <- boot(data = as.data.frame(for.woo), statistic = bs.sor, R = 2000, 
                   formula = (1 - Labd) ~ ipcc)

wl.jmixres <- boot(data = as.data.frame(mix.shr), statistic = bs.jac, R = 2000, 
                   formula = (1 - Jabd) ~ ipcc)
wl.smixres <- boot(data = as.data.frame(mix.shr), statistic = bs.sor, R = 2000, 
                   formula = (1 - Labd) ~ ipcc)

wl.jtalres <- boot(data = as.data.frame(tal.shr), statistic = bs.jac, R = 2000, 
                   formula = (1 - Jabd) ~ ipcc)
wl.stalres <- boot(data = as.data.frame(tal.shr), statistic = bs.sor, R = 2000, 
                   formula = (1 - Labd) ~ ipcc)

wl.jshrres <- boot(data = as.data.frame(shr.shr), statistic = bs.jac, R = 2000, 
                   formula = (1 - Jabd) ~ ipcc)
wl.sshrres <- boot(data = as.data.frame(shr.shr), statistic = bs.sor, R = 2000, 
                   formula = (1 - Labd) ~ ipcc)

wl.jmeares <- boot(data = as.data.frame(agr.mea), statistic = bs.jac, R = 2000, 
                   formula = (1 - Jabd) ~ ipcc)
wl.smeares <- boot(data = as.data.frame(agr.mea), statistic = bs.sor, R = 2000, 
                   formula = (1 - Labd) ~ ipcc)

Between Landscapes

# bootstrapping with 2000 replications
bl.jwoores <- boot(data = as.data.frame(woo), statistic = bs.jac, R = 2000, 
                   formula = (1 - Jabd) ~ ipcc)
bl.swoores <- boot(data = as.data.frame(woo), statistic = bs.sor, R = 2000, 
                   formula = (1 - Labd) ~ ipcc)

bl.jmixres <- boot(data = as.data.frame(mix), statistic = bs.jac, R = 2000, 
                   formula = (1 - Jabd) ~ ipcc)
bl.smixres <- boot(data = as.data.frame(mix), statistic = bs.sor, R = 2000, 
                   formula = (1 - Labd) ~ ipcc)

bl.jtalres <- boot(data = as.data.frame(tal), statistic = bs.jac, R = 2000, 
                   formula = (1 - Jabd) ~ ipcc)
bl.stalres <- boot(data = as.data.frame(tal), statistic = bs.sor, R = 2000, 
                   formula = (1 - Labd) ~ ipcc)

bl.jshrres <- boot(data = as.data.frame(shr), statistic = bs.jac, R = 2000, 
                   formula = (1 - Jabd) ~ ipcc)
bl.sshrres <- boot(data = as.data.frame(shr), statistic = bs.sor, R = 2000, 
                   formula = (1 - Labd) ~ ipcc)

bl.jmeares <- boot(data = as.data.frame(mea), statistic = bs.jac, R = 2000, 
                   formula = (1 - Jabd) ~ ipcc)
bl.smeares <- boot(data = as.data.frame(mea), statistic = bs.sor, R = 2000, 
                   formula = (1 - Labd) ~ ipcc)

All Biotopes

# bootstrapping with 2000 replications
al.jforres <- boot(data = as.data.frame(woo.bio), statistic = bs.jac, R = 2000, 
                   formula = (1 - Jabd) ~ ipcc)
al.sforres <- boot(data = as.data.frame(woo.bio), statistic = bs.sor, R = 2000, 
                   formula = (1 - Labd) ~ ipcc)

al.jmixres <- boot(data = as.data.frame(mix.bio), statistic = bs.jac, R = 2000, 
                   formula = (1 - Jabd) ~ ipcc)
al.smixres <- boot(data = as.data.frame(mix.bio), statistic = bs.sor, R = 2000, 
                   formula = (1 - Labd) ~ ipcc)

al.jtalres <- boot(data = as.data.frame(tal.bio), statistic = bs.jac, R = 2000, 
                   formula = (1 - Jabd) ~ ipcc)
al.stalres <- boot(data = as.data.frame(tal.bio), statistic = bs.sor, R = 2000, 
                   formula = (1 - Labd) ~ ipcc)

al.jshrres <- boot(data = as.data.frame(sho.bio), statistic = bs.jac, R = 2000, 
                   formula = (1 - Jabd) ~ ipcc)
al.sshrres <- boot(data = as.data.frame(sho.bio), statistic = bs.sor, R = 2000, 
                   formula = (1 - Labd) ~ ipcc)

al.jmeares <- boot(data = as.data.frame(mea.bio), statistic = bs.jac, R = 2000, 
                   formula = (1 - Jabd) ~ ipcc)
al.smeares <- boot(data = as.data.frame(mea.bio), statistic = bs.sor, R = 2000, 
                   formula = (1 - Labd) ~ ipcc)

Confidence Interval

Within Landscape - get 95% confidence intervals

Jaccard Estimator

wl.jforres.ci <- ci.boot(wl.jforres)
wl.jmixres.ci <- ci.boot(wl.jmixres)
wl.jtalres.ci <- ci.boot(wl.jtalres)
wl.jshrres.ci <- ci.boot(wl.jshrres)

## Warning: extreme order statistics used as endpoints

wl.jmeares.ci <- ci.boot(wl.jmeares)

Sorensen Estimator

wl.sforres.ci <- ci.boot(wl.sforres)
wl.smixres.ci <- ci.boot(wl.smixres)
wl.stalres.ci <- ci.boot(wl.stalres)
wl.sshrres.ci <- ci.boot(wl.sshrres)
wl.smeares.ci <- ci.boot(wl.smeares)

Between Landscapes - get 95% confidence intervals

Jaccard Estimator

bl.jforres.ci <- ci.boot(bl.jwoores)
bl.jmixres.ci <- ci.boot(bl.jmixres)
bl.jtalres.ci <- ci.boot(bl.jtalres)
bl.jshrres.ci <- ci.boot(bl.jshrres)
bl.jmeares.ci <- ci.boot(bl.jmeares)

Sorensen Estimator

bl.sforres.ci <- ci.boot(bl.swoores)
bl.smixres.ci <- ci.boot(bl.smixres)
bl.stalres.ci <- ci.boot(bl.stalres)
bl.sshrres.ci <- ci.boot(bl.sshrres)
bl.smeares.ci <- ci.boot(bl.smeares)

All Biotopes - get 95% confidence intervals

Jaccard Estimator

al.jforres.ci <- ci.boot(al.jforres)
al.jmixres.ci <- ci.boot(al.jmixres)
al.jtalres.ci <- ci.boot(al.jtalres)
al.jshrres.ci <- ci.boot(al.jshrres)
al.jmeares.ci <- ci.boot(al.jmeares)

Sorensen Estimator

al.sforres.ci <- ci.boot(al.sforres)
al.smixres.ci <- ci.boot(al.smixres)
al.stalres.ci <- ci.boot(al.stalres)
al.sshrres.ci <- ci.boot(al.sshrres)
al.smeares.ci <- ci.boot(al.smeares)

Results 1

Jaccard

Intercept

par(mfrow = c(1, 3))
interc.plot3(wl.jforres.ci, wl.jmixres.ci, wl.jmeares.ci)
title("Within Landscape", cex = 1.5)
interc.plot3(bl.jforres.ci, bl.jmixres.ci, bl.jmeares.ci)
title("Between Landscapes", cex = 1.5)
interc.plot3(al.jforres.ci, al.jmixres.ci, al.jmeares.ci)
title("All Biotopes", cex = 1.5)

plot of chunk interc3_jac

Dissimilarity

par(mfrow = c(1, 3))
dissim.plot3(wl.jforres.ci, wl.jmixres.ci, wl.jmeares.ci)
title("Within Landscape", cex = 1.5)
dissim.plot3(bl.jforres.ci, bl.jmixres.ci, bl.jmeares.ci)
title("Between Landscapes", cex = 1.5)
dissim.plot3(al.jforres.ci, al.jmixres.ci, al.jmeares.ci)
title("All Biotopes", cex = 1.5)

plot of chunk dissim3_jac

Slope

par(mfrow = c(1, 3))
slope.plot3(wl.jforres.ci, wl.jmixres.ci, wl.jmeares.ci)
title("Within Landscape", cex = 1.5)
slope.plot3(bl.jforres.ci, bl.jmixres.ci, bl.jmeares.ci)
title("Between Landscapes", cex = 1.5)
slope.plot3(al.jforres.ci, al.jmixres.ci, al.jmeares.ci)
title("All Biotopes", cex = 1.5)

plot of chunk slope3_jac

R-Square

par(mfrow = c(1, 3))
rsquar.plot3(wl.jforres.ci, wl.jmixres.ci, wl.jmeares.ci)
title("Within Landscape", cex = 1.5)
rsquar.plot3(bl.jforres.ci, bl.jmixres.ci, bl.jmeares.ci)
title("Between Landscapes", cex = 1.5)
rsquar.plot3(al.jforres.ci, al.jmixres.ci, al.jmeares.ci)
title("All Biotopes", cex = 1.5)

plot of chunk r-quar3_jac

Sorensen

Intercept

par(mfrow = c(1, 3))
interc.plot3(wl.sforres.ci, wl.smixres.ci, wl.smeares.ci)
title("Within Landscape", cex = 1.5)
interc.plot3(bl.sforres.ci, bl.smixres.ci, bl.smeares.ci)
title("Between Landscapes", cex = 1.5)
interc.plot3(al.sforres.ci, al.smixres.ci, al.smeares.ci)
title("All Biotopes", cex = 1.5)

plot of chunk interc3_sor

Dissimilarity

par(mfrow = c(1, 3))
dissim.plot3(wl.sforres.ci, wl.smixres.ci, wl.smeares.ci)
title("Within Landscape", cex = 1.5)
dissim.plot3(bl.sforres.ci, bl.smixres.ci, bl.smeares.ci)
title("Between Landscapes", cex = 1.5)
dissim.plot3(al.sforres.ci, al.smixres.ci, al.smeares.ci)
title("All Biotopes", cex = 1.5)

plot of chunk dissim3_sor

Slope

par(mfrow = c(1, 3))
slope.plot3(wl.sforres.ci, wl.smixres.ci, wl.smeares.ci)
title("Within Landscape", cex = 1.5)
slope.plot3(bl.sforres.ci, bl.smixres.ci, bl.smeares.ci)
title("Between Landscapes", cex = 1.5)
slope.plot3(al.sforres.ci, al.smixres.ci, al.smeares.ci)
title("All Biotopes", cex = 1.5)

plot of chunk slope3_sor

R-Square

par(mfrow = c(1, 3))
rsquar.plot3(wl.sforres.ci, wl.smixres.ci, wl.smeares.ci)
title("Within Landscape", cex = 1.5)
rsquar.plot3(bl.sforres.ci, bl.smixres.ci, bl.smeares.ci)
title("Between Landscapes", cex = 1.5)
rsquar.plot3(al.sforres.ci, al.smixres.ci, al.smeares.ci)
title("All Biotopes", cex = 1.5)

plot of chunk r-squar3_sor

Results 2

Jaccard

Intercept

par(mfrow = c(1, 3))
interc.plot4(wl.jforres.ci, wl.jtalres.ci, wl.jshrres.ci, wl.jmeares.ci)
title("Within Landscape", cex = 1.5)
interc.plot4(bl.jforres.ci, bl.jtalres.ci, bl.jshrres.ci, bl.jmeares.ci)
title("Between Landscapes", cex = 1.5)
interc.plot4(al.jforres.ci, al.jtalres.ci, al.jshrres.ci, al.jmeares.ci)
title("All Biotopes", cex = 1.5)

plot of chunk interc4_jac

Dissimilarity

par(mfrow = c(1, 3))
dissim.plot4(wl.jforres.ci, wl.jtalres.ci, wl.jshrres.ci, wl.jmeares.ci)
title("Within Landscape", cex = 1.5)
dissim.plot4(bl.jforres.ci, bl.jtalres.ci, bl.jshrres.ci, bl.jmeares.ci)
title("Between Landscapes", cex = 1.5)
dissim.plot4(al.jforres.ci, al.jtalres.ci, al.jshrres.ci, al.jmeares.ci)
title("All Biotopes", cex = 1.5)

plot of chunk dissim4_jac

Slope

par(mfrow = c(1, 3))
slope.plot4(wl.jforres.ci, wl.jtalres.ci, wl.jshrres.ci, wl.jmeares.ci)
title("Within Landscape", cex = 1.5)
slope.plot4(bl.jforres.ci, bl.jtalres.ci, bl.jshrres.ci, bl.jmeares.ci)
title("Between Landscapes", cex = 1.5)
slope.plot4(al.jforres.ci, al.jtalres.ci, al.jshrres.ci, al.jmeares.ci)
title("All Biotopes", cex = 1.5)

plot of chunk slope4_jac

R-Square

par(mfrow = c(1, 3))
rsquar.plot4(wl.jforres.ci, wl.jtalres.ci, wl.jshrres.ci, wl.jmeares.ci)
title("Within Landscape", cex = 1.5)
rsquar.plot4(bl.jforres.ci, bl.jtalres.ci, bl.jshrres.ci, bl.jmeares.ci)
title("Between Landscapes", cex = 1.5)
rsquar.plot4(al.jforres.ci, al.jtalres.ci, al.jshrres.ci, al.jmeares.ci)
title("All Biotopes", cex = 1.5)

plot of chunk r-squar4_jac

Sorensen

Intercept

par(mfrow = c(1, 3))
interc.plot4(wl.sforres.ci, wl.stalres.ci, wl.sshrres.ci, wl.smeares.ci)
title("Within Landscape", cex = 1.5)
interc.plot4(bl.sforres.ci, bl.stalres.ci, bl.sshrres.ci, bl.smeares.ci)
title("Between Landscapes", cex = 1.5)
interc.plot4(al.sforres.ci, al.stalres.ci, al.sshrres.ci, al.smeares.ci)
title("All Biotopes", cex = 1.5)

plot of chunk interc4_sor

Dissimilarity

par(mfrow = c(1, 3))
dissim.plot4(wl.sforres.ci, wl.stalres.ci, wl.sshrres.ci, wl.smeares.ci)
title("Within Landscape", cex = 1.5)
dissim.plot4(bl.sforres.ci, bl.stalres.ci, bl.sshrres.ci, bl.smeares.ci)
title("Between Landscapes", cex = 1.5)
dissim.plot4(al.sforres.ci, al.stalres.ci, al.sshrres.ci, al.smeares.ci)

plot of chunk dissim4_sor

Slope

par(mfrow = c(1, 3))
slope.plot4(wl.sforres.ci, wl.stalres.ci, wl.sshrres.ci, wl.smeares.ci)
title("Within Landscape", cex = 1.5)
slope.plot4(bl.sforres.ci, bl.stalres.ci, bl.sshrres.ci, bl.smeares.ci)
title("Between Landscapes", cex = 1.5)
slope.plot4(al.sforres.ci, al.stalres.ci, al.sshrres.ci, al.smeares.ci)
title("All Biotopes", cex = 1.5)

R-Square

par(mfrow = c(1, 3))
rsquar.plot4(wl.sforres.ci, wl.stalres.ci, wl.sshrres.ci, wl.smeares.ci)
title("Within Landscape", cex = 1.5)
rsquar.plot4(bl.sforres.ci, bl.stalres.ci, bl.sshrres.ci, bl.smeares.ci)
title("Between Landscapes", cex = 1.5)
rsquar.plot4(al.sforres.ci, al.stalres.ci, al.sshrres.ci, al.smeares.ci)
title("All Biotopes", cex = 1.5)

plot of chunk r-squar4_sor

Tables

Jaccard

Within landscape

Between landscapes

All biotopes

Sorensen

Within landscape

Between landscapes

All biotopes

Moth

Murilo Miranda

9 Sep 2014

Load Data and Codes

Distances

Data

Exploratory Analysis

FIRST APROARCH

Within landscape

Tables

SECOND APPROACH

Bootstrapping 1

Mean Dissimilarity for each scale

Bootstrapping 2

Within Landscape

Between Landscapes

All Biotopes

Within Landscape - get 95% confidence intervals

Between Landscapes - get 95% confidence intervals

All Biotopes - get 95% confidence intervals

Results 1

Results 2

Tables

Jaccard

Sorensen